The Creative AI in Retail


Artificial Intelligence has been adopted across every industry in recent years. From health care to aviation, the power of the machines is being unleashed to perform time-critical tasks, make informed decisions and make our lives better every day.

The Retail Industry is no exception. Today, when you add a pair of jeans to your cart on your favorite online retailer, you are immediately shown a wide variety of shoes, tops and other items as a recommendation, which will complement your fashion sense. These recommended products are carefully handpicked by the system, which has analyzed your previous purchases and other users’ shopping trends. All of this is possible with the help of machine learning which trains the system with tons of data and teaches it to infer patterns lying underneath the chaos and hence enhances the system’s deductive skills.

Machine learning sounds more like it tries to mimic the left side of the human brain, which is responsible for logic, reasoning and analytics. Can we somehow use machine learning to mimic the right side of the brain, which is responsible for intuition, imagination and most importantly, creativity?

The answer lies in one of the classes of machine learning known as GAN (Generative adversarial network). A GAN can generate photographs with such realistic features and accuracy that you may get confused whether it’s the work of a machine or a real person. To start with, a computer is trained with a known set of images, such as a bunch of paintings. The GAN then aids the computer in recognizing the building blocks of those paintings, and then picks up those blocks at random and sews them together to generate a completely new piece of artwork! We may not have a digital Da Vinci yet, the applications of GAN are becoming widespread, and is especially taking the retail industry by storm.

The key to an engaging shopping experience is to help the customer visualize themselves in the apparel they wish to purchase. GANs take this experience to a whole new level. If people have trouble in deciding whether or not to buy a shirt, they can simply upload a picture of themselves, and the computer will generate a realistic image of theirs wearing the shirt. The computer can also go ahead in creating images of a person wearing the shirt, standing in different poses, irrespective of the pose of the person in the image provided as input. So, imagine a new recommendation system which doesn’t show only the apparel, but is also creative and intuitive enough to generate the perfect image of yours, in that apparel.

GANs can also take head-on with Augmented reality (AR) systems in the future and emerge victorious! Cause GANs don’t have limitations of AR systems in terms of lighting conditions and device form factors.

GANs may also land up in being used in Retail therapy, which is shopping with the motive of improving the buyer’s mood. Some people often resort to retail therapy in their times of depression, relying on the notion that a new pair of clothes will change their outlook towards lives. This notion doesn’t come cheap, and sometimes people spend more than they originally intended to. Since the primary purpose of retail therapy is to give a morale boost to a person, the images generated by GANs will not only make a person a conscious buyer but also make their lives better by re-instilling confidence in themselves.

How can we help?